Validity test Reliability test

Apin Hidayat, 2013 The Use Of Songs In Teaching Students Listening Ability Universitas Pendidikan Indonesia | repository.upi.edu

3.5.2.1 Validity test

To analyze validity of the test, this research was used external validity analysis by correlating the students’ score of listening test with students’ summative score of English which were taken from their mid-term report book. To correlate them, the formula of Pearson Product Moment Correlation Coefficient Values is used. Where r is Pearson Product Moment Correlation Coefficient Values, N is number of students who are analyzed, x is students’ listening score first variable and y is students’ summative score second variable. Afterwards, every item on the instrument was also analyzed by using Pearson Product Moment Correlation Coefficient Values gained by distributing the data to statistical program; Anates v.4 to determine correlation index for every items. This analysis was conducted through correlating every single item of instrument x with total score of instrument y . Next, the r is interpreted as follows: Apin Hidayat, 2013 The Use Of Songs In Teaching Students Listening Ability Universitas Pendidikan Indonesia | repository.upi.edu Table 3.3 Index of Validity Level Coefficient Interval r Validity Level 0.800 – 1.000 Excellent Ex 0.600 – 0.799 Good Gd 0.400 – 0.599 Satisfying St 0.200 – 0.399 Poor Pr 0.000 – 0.199 Very Poor Vp Adopted from Riduwan 2004:110

3.5.2.2 Reliability test

This research used KUDER RICHARDSON -20 KR-20 method to analyze reliability of instrument. The purpose of using this method is to gain a higher reliability value, since the result of KR-20 tends to give a higher value than the other methods such as KR-21, Anova Hoyt, and Alpha . Arikunto 1999. Every single right answer is valued 1 and every wrong answer is 0. Apin Hidayat, 2013 The Use Of Songs In Teaching Students Listening Ability Universitas Pendidikan Indonesia | repository.upi.edu KUDER RICHARDSON -20 KR-20 formula: r 11 = Internal reliability coefficient for all items K = sum of question item P = proportion of subjects who answer right q = proportion of subjects who answer wrong q = 1 – p s 2 = variance total Afterward, the value of r 11 is compared with index of Pearson Product Moment Correlation Coefficient Values to see whether the value is reliable or not.

3.5.3 Treatments